Efficient Backdoor Attacks for Deep Neural Networks in Real-world Scenarios

TL;DR

This paper proposes a new, efficient backdoor attack method for deep neural networks that works effectively even when attackers have limited access to training data from multiple sources, outperforming previous methods.

Contribution

Introduces a novel data-constrained backdoor attack approach using CLIP-based techniques, addressing limitations of existing methods under realistic data access restrictions.

Findings

Achieves over 100% improvement in attack success rate in constrained scenarios.

Effectively suppresses benign features and augments poisoning features.

Demonstrates robustness of the proposed method in real-world data collection scenarios.

Abstract

Recent deep neural networks (DNNs) have came to rely on vast amounts of training data, providing an opportunity for malicious attackers to exploit and contaminate the data to carry out backdoor attacks. However, existing backdoor attack methods make unrealistic assumptions, assuming that all training data comes from a single source and that attackers have full access to the training data. In this paper, we introduce a more realistic attack scenario where victims collect data from multiple sources, and attackers cannot access the complete training data. We refer to this scenario as data-constrained backdoor attacks. In such cases, previous attack methods suffer from severe efficiency degradation due to the entanglement between benign and poisoning features during the backdoor injection process. To tackle this problem, we introduce three CLIP-based technologies from two distinct streams: Clean Feature Suppression and Poisoning Feature Augmentation.effective solution for data-constrained backdoor attacks. The results demonstrate remarkable improvements, with some settings achieving over 100% improvement compared to existing attacks in data-constrained scenarios. Code is available at https://github.com/sunh1113/Efficient-backdoor-attacks-for-deep-neural-networks-in-real-world-scenarios